Is AI making it up?
The more specific an answer sounds, the more it feels checked. It isn’t.
Every question in FutureMinds is written twice. So is this lesson.
Would you trust that answer?
There is no forest, no crystals and no Shadow Boss. An AI that has never played the game can still describe one, because it has read thousands of pages about games like it and knows the shape a secret-character guide takes.
Nothing in the answer is marked as a guess. That is the problem — it reads exactly like a real walkthrough.
Would you trust that answer?
That is the exact shape of a real citation, which is why it convinces. An AI can produce the shape without the substance: plausible name, plausible institution, plausible year, and no source behind any of it.
Invented references are the most common way this goes wrong in schoolwork, and they are hard to spot precisely because they look correct.
A detailed answer is not a checked answer.
In the app today: AI Galaxy · Know It or Guess It? asks a child to rule on each line of a reply — did it know that, or guess it?
One question catches all four.
Don’t ask only “What does AI say?”
Ask “How does AI know?”
Your child can practise this
FutureMinds teaches children 9–16 how AI works — through quests, not lectures. Launching 15 September.
Get FutureMinds at launch →